 Hey folks, it's T Tuesday Sharp-eyed viewers will have noticed like David Kiersey did last month that I did not actually have any goals for this update It didn't even say that I was gonna be here, but here we are The main thing of course was getting the living computation theory of everything video out Here's what I've been thinking about Come on, there we go. So I want to talk about the video a little bit Take a moment to look back on the challenge that's just done And then stick myself with a new one. We'll see how it goes So living computation theory of everything, you know On the negative side, you know, it's still too hard to follow It's still preaching to the choir where the choir consists primarily of me But on the other hand, I've been trying to say this stuff for a while and and this current version is Is is definitely the most coherent one yet and the defined defendant attack where I say things Objections that people might raise and then say how I would respond to them I think it actually helps quite a bit for people to get a handle on Where what I'm saying actually is and one of the things that I kind of like about the way that it all finished up was How it really centers humans humanism and humanity Rather than, you know, the cold world of protons and and and sub-maisons or whatever it is Skip the part about the demo at the moment, but it's both good and bad Okay So the first half of this year was supposed to be this multicellular challenge It didn't start out as a multicellular challenge It started out as a digital replicator And this was our master plan that we looked at a many times the mother Little diamond with some code in it. It would grow bigger or move the code around it would copy the code and eventually it would cut And then we would end up with with two Completely legitimate mothers that would go on to do the same thing and that is the recipe for unicellular life There's nothing saying in this picture saying that these two things are supposed to team up and do something But we actually did get to that So multicellular challenge challenge met zero to multicellular in six months I mean It's not too bad, you know for a small project and you know, there are Many other pieces of work that have been done over the years and decades that have much in similarities To what we did here particularly but in its details, you know, it's it's absolutely really the first of its kind in the world and Okay, that's that's a little bit of something. I personally, you know learned a ton And I thought about all of these things that you know, when you don't have global architectural synchronization You need to have self organizing local synchronization and how you do that There's lots of different choices and the ring oscillators and the gradients and all that stuff But going through the diamond replicators and then the multicellular stuff, you know, really sort of made it Built it in to my thinking it's it's not something that I have to think to apply It's just you know, okay. How big is this thing going to be? So how much delays are we going to be talking about? What kind of scaling do we want to do and you know? It is pretty you know, I've gotten awful used to it, but it is pretty cool to watch on the other hand There's way too much physics and not nearly enough chemistry and biology in the model for my taste You know, there's just enough to make it, you know, non-trivially biological But the bigger problem is that the model itself the whole idea is too thin from its smallest Component to its biggest component is not thick enough You know a single atom an Entire diamond which is meant to be a replicator, you know like a cell or something like that is only a couple of thousand atoms So so that's like three orders of magnitude from the smallest thing to the thing that's supposed to be a replicator Whereas, you know in our actual real physics, you know the absolute teeniest replicators are in billions of atoms and then on up and Because we are so thin because we have so little space Between the out of the elementary particle and the creature There's just no room to store, you know giant sequences of DNA You can do that when you have billions of atoms in a single replicator You can't do it when you have thousands of atoms most of which you're spoken for In addition, there are still lots of fragilities in the code hc3 hard cell 3 the fundamental grid that made all of the mobile Diamond stuff possible still has deadlocks in it and rah rah rah rah if the code is hard to work with it's slow to build it's slow to run and I mean I did work. I did push pretty hard on the diamonds over the you know last four months or whatever it was over 2023 really and I'm a little sick of it. I want to do something else. So New challenge I want to start by showing a little clip from living computation theory of everything and use that to set up the new challenge Now How are we doing on our second implementation in fact in my mind in my heart the t2 tile project is Working on a second implementation, but you know How far is the t2 tile project from being able to respond to code like this and do anything It's a zillion miles away. And so, you know, that's why it's an IOU So that's it The t2 tile the MFM has a lot of good stuff, it's spatially distributed it doesn't this presumed determinism so on and so forth It's indefinitely scalable But there's no way anytime soon. We're gonna be able to go from atoms to on up the scale and Get as far as you know us just just no way So one idea is Let's move the model to a different area of space time. Let's regroup And in particular what if the t2 tile matrix the entire matrix However big it is currently 76 tiles and stuff was all meant to be just a little patch of Brain and the body and the connection between the body and the brain is Outside the grid outside the matrix. That's what I want to go for t2 matrix meets world So, you know, I tell people they asked me what's the killer app for this? When am I gonna be able to run an excel on? The t2 tiles and it's like, you know, no time soon in my response it traditionally is Robust system control robust real-time system control is what the t2 matrix style of computation the bottom up Distributed the robust is good for because you really if if the system that's being controlled is The life and death of that system, you know, then you really do need it to be robust you really don't want to be thinking about penny-binging efficiency if the a failure of the system takes out the whole thing so But of course The t2's are so slow and so forth. We can't connect to the real world So the idea is we're gonna have some kind of simulator over here. That's gonna simulate something But we have to figure out how to connect the t2 matrix to that something and we'll be able to crank down the speed On the simulated world to match the incredibly slow speed on the t2 tiles and we'll see what happens So tons of questions that I am mostly am not ready to give good answers to yet But figuring that out is part of the task of the challenge, you know What is the system that's being controlled? What is the world that the system is in how to connect the t2 matrix to the world and how to deal with real time and so forth I'm gonna focus on how to connect t2 matrix to the real world because that's kind of logically prior before we can do anything We have to figure out how we're gonna get IO in and out of the matrix So one possibility is right, you know, so here's a t2 tile this is the West edge of the t2 tile and it's got a an ethernet port Right there and you can only get to the ethernet port if the West edge of the tile isn't covered up by another tile But we could connect to all of the Western tiles We could plug in an ethernet cable put them all into a hub and then network our way to the computer. That's running Whatever the world and the system that's being controlled is and you know, we could do this today At least in a small version. I need to get a bigger switch later on But I'm really not satisfied with it. It's you know one edge Seems very arbitrary Originally, I'd been thinking that you know, we'd have an interface all around the edge of The matrix but when I started thinking about it as being like a patch of brain I really wanted to talk to the whole thing I mean just like we have the spinal cord coming in and hitting Cortex and the surface of the brain and so forth and there is a possibility of how to do that So in addition to the ethernet port there is a serial port You are On each tile that is accessible from the front So even if we have the things all tiled up in principle, we can get at a serial port on each you on each T2 That's what this sketch is drawing here These are serial cables and then we have something magic happens and then all of those serial ports connect to the computer That's running the simulation and so I started figuring out, you know, what the heck could go in here I mean, so like here is four serial ports going into USB and so the idea is like, you know, we can get like You know 30 or 40 of these or something or 25 anyway You get to like a hundred and then we would have to get you USB hubs daisy chain together feeding them together this I cannot I cannot imagine letting myself try to do that Another approach is an actual rack server thing. This thing is got 32 serial ports that it puts them all down onto one ethernet port like that Cost-effective serial concentrator. So it would take like three of these to Get enough serial ports to talk to all of the T2 tops. No, maybe we don't have to talk to all of them But I'd like to And you know, I could go to the hardware dealer people, you know, Mausers one of the ones I'm sort of comfortable and I could have bought to buy it today And they're $2,600 a pop so cost-effective in this land means less than a hundred dollars per port You know if this was really a great solution then, you know, maybe the money would be Tolerable but there are other issues that I have I'm not really super happy about it Well, then there's this thing the AM 3 3 3 5x supports up to six additional you arts. Well, so what who cares? Well, because the AM 3 AM 3 3 5x is what's inside the T2 tiles Now we can't use those you arts on the T2 tiles because the pins are already doing other stuff But if we this is a beagle bone green This is the processor that's inside of every T2 tile underneath this Heat sink here is an AM 3 3 5x So we could maybe do a new little bit of circuitry a new little bit of programming and get a Have one of these that would have six and now it's possible that there could mean have four others and End up actually it says six additional that in principle that could be on top of four So one of these might be able to take ten serial ports again It wouldn't necessarily be able to take them all at super high speed But this is the good thing about this particular task everything is so slow that we don't need to max out the serial port speeds Because everything is going to be so far clocked down to compensate for the slow average event rate of the T2 tiles so A dozen of these maybe Maybe ten could cover the whole thing. I've got that sitting in the stock right now that you know Beagle bone greens that haven't yet been attached to T2 tiles and so forth So this is the idea BBG beagle bone green I wrote it is taking six you arts each so we would have like a dozen of these or something They would each have an ethernet cable that would go to a switch and then would go in the serial port concentrator I'm gonna go for this. I'm gonna try Who will see so This is what we're talking about it's the T2 matrix brain challenge not human brain Not even you know Mouse brain it's some tiny tiny tiny little Stimulus response just the the barest minimum sort of thing And I've made up a schedule. This was the hardest part of the whole thing was to actually pin myself down So we are already right now is matrix brain challenge t-minus six counting down by t Tuesday updates So there is not gonna be a update in November because once again I'm going to do nano nano-remo the national novel writing month in this case It's not even close to trying to write a novel instead. It's gonna be to bring out the collected works of von Joy Manon Which will include the science fiction story search quiet wake that I was trying to sell the other months They're the months and years back And we'll take it from there and we'll cover all this stuff Brightenburg vehicle sensory motor homunculus as we go forward and That's the plan serial port concentrator demo collected works out have a lot of fun Thanks so much for stopping in I hope to see you in December